The association between BMI and cognition in India: data from the Longitudinal Aging Study in India (LASI).

BACKGROUND

High body-mass index (BMI) is an established risk factor for late-life cognitive impairment and dementia, but most evidence comes from high-income contexts. Existing evidence from cross-sectional data in low- and middle-income settings is inconsistent, and many studies do not adequately address potential sources of bias.

METHODS

We used data from Wave 1 of the Longitudinal Aging Study in India (LASI) (analytic N = 56,753) to estimate the association between BMI categories and cognitive functioning among older adults aged 45 + years using survey-weighted linear regression models stratified by gender and controlling for potential confounders including demographic factors, socio-economic status (SES) characteristics, and health-related behaviors. To probe potential sources of bias, including residual confounding and reverse causation, we used weighting and trimming methods, sample restriction, and explored effect modification.

RESULTS

In fully adjusted models, relative to normal BMI underweight BMI was associated with lower cognitive scores (Men: -0.16 SD difference, 95% CI -0.18, -0.13; Women: -0.12 SD, -0.15, -0.10). Overweight and obesity were associated with higher cognitive scores in both men (overweight: 0.09; 0.07, 0.12, obese: 0.10; 0.05, 0.15) and women (overweight: 0.09; 0.07-0.12, obese: 0.12; 0.08-0.15). Estimates were similar after weighting and trimming but were attenuated after excluding those with low cognition (≥1 SD below the mean relative to those with similar demographic characteristics). Positive associations between overweight and obese BMI and cognition were attenuated or null in those living in urban settings and those with higher levels of educational attainment.

CONCLUSIONS

Underweight BMI is a risk factor for poor cognitive outcomes in adults 45 years and older and may be indicative of poor nutritional status and life-course disadvantage in India. In tandem with existing literature, supplemental analyses and effect modification results indicate that unmeasured confounding and reverse causation may explain the observed positive associations between overweight and obese BMI and cognitive functioning from cross-sectional studies in low- and middle-income settings. Future data with longitudinal follow-up will be helpful to further disentangle biases.